Most LLMs return text; teams need structured decisions. This marketplace treats product practice as first-class agent artifacts: PM frameworks (assumption mapping, Opportunity Solution Trees, PRDs, OKRs, GTM) are packaged as individual skills and chained commands so assistants produce reviewable, repeatable PM work—not just prose.
What Sets It Apart
- Skill-first design that mirrors practiced PM methods — each skill encodes a named framework (e.g., OST, prioritization, experiment design). So what: prompts become procedural workflows that preserve intent and steps rather than ad-hoc replies.
- Installable plugins + slash commands for Claude Code/Cowork (and skill files that port to other assistants). So what: non-developers get one-click workflows (/discover, /plan-launch, /write-prd) while power users can import skills into CLIs or other tooling.
- Focus on end-to-end PM lifecycle, including an "AI shipping" kit. So what: beyond ideation and strategy, it supplies artifacts and audits (document-app, ship-check, derive-tests) aimed at making AI-built code reviewable and auditable.
- Practical, framework-driven outputs (interview scripts, PRDs, test maps, release notes). So what: reduces the manual gap between product thinking and ship-ready artifacts.
Who it fits & tradeoffs
Great fit if you: are a product manager or small PM-led team wanting repeatable, framework-driven workflows inside an AI assistant; use Claude Code/Cowork or are comfortable importing skill folders into other assistants; want artifactable outputs (PRDs, OSTs, experiment plans) rather than freeform brainstorming.
Look elsewhere if you: need a vendor-agnostic GUI product for enterprise-scale governance, require deep native integrations with non-LLM tooling out of the box, or expect the repo to replace core PM tooling (Jira, product analytics) rather than augment workflows. The marketplace depends on the host assistant's runtime and behavior—some commands are Claude-specific and may require adaptation for other platforms.
